SETH SWIRSKYv. MARIAH CAREY
376 F.3d 841 (9th Cir. 2004)
CANBY, Circuit Judge:
The plaintiffs, Seth Swirsky and Warryn Campbell, brought this action in district court, alleging that a song produced by the defendants infringed the plaintiffs' copyright in the song, "One of Those Love Songs." The defendants moved for summary judgment, contending that the plaintiffs' evidence failed to meet this circuit's threshold "extrinsic test" for substantial similarity of works. The district court granted the motion, holding that the plaintiffs' expert had failed to show by external, objective criteria that the two songs shared a similarity of ideas and expression. Plaintiffs appeal. We conclude that the plaintiffs' expert's evidence was sufficient to present a triable issue of the extrinsic similarity of the two songs, and that the district court's ruling to the contrary was based on too mechanical an application of the extrinsic test to these musical compositions. We also conclude that the district court erred in ruling portions of plaintiffs' song to be unprotectable by copyright as a matter of law. We accordingly reverse the summary judgment.
Factual Background
This case concerns the alleged similarity between the choruses of two popular and contemporary rhythm and blues ("R&B") songs: plaintiffs' "One of Those Love Songs" ("One") and Mariah Carey's "Thank God I Found You" ("Thank God"). One was jointly composed by plaintiffs Seth Swirsky and Warryn Campbell (collectively "Swirsky") in 1997. Pursuant to a licensing agreement, One was recorded [*844] by the musical group Xscape and released in May 1998 on Xscape's album "Traces of My Lipstick." Thank God was composed by defendants Carey, James Harris III, and Terry Lewis in 1999 and was released on Carey's album "Rainbow" in November 1999.
One and Thank God have generally dissimilar lyrics and verse melodies, but they share an allegedly similar chorus that Swirsky claims as an infringement of One's copyright.[1] Swirsky filed this action in district court against Carey, Harris, Lewis, and a number of music companies that had financial interests in Thank God (collectively "Carey") for copyright infringement and related claims. The defendants moved for summary judgment, contending that Swirsky had failed to present a triable issue on the required first, or "extrinsic," part of our circuit's two-part test for the establishment of substantial similarity necessary to sustain a claim of copyright infringement. The defendants also contended that portions of One were not protectable by copyright. The district court agreed with both contentions and granted summary judgment to Carey. Swirsky moved for reconsideration, which the district court denied. This appeal followed.
Substantial Similarity
We review de novo the district court's grant of summary judgment. See Smith v. Jackson, 84 F.3d 1213, 1218 (9th Cir. 1996). We may uphold the summary judgment only if we find that "no reasonable juror could find substantial similarity of ideas and expression [between One and Thank God], viewing the evidence in the light most favorable to the nonmoving party." Id. If Swirsky presented "indicia of a sufficient disagreement concerning the substantial similarity of [the] two works," then the case must be submitted to a trier of fact.
To establish a successful copyright infringement claim, Swirsky must show that (1) he owns the copyright in One and (2) Carey copied protected elements of One. See Rice v. Fox Broad. Co., 330 F.3d 1170, 1174 (9th Cir. 2003) ( Rice I); For purposes of summary judgment, Carey conceded that Swirsky owns a valid copyright in One. The element of copying is rarely the subject of direct evidence; Swirsky may establish copying by showing that Carey had access to One and that Thank God was substantially similar to One in One's protected elements. See Smith, 84 F.3d at 1218; Metcalf v. Bochco, 294 F.3d 1069, 1072 (9th Cir. 2002). Where a high degree of access is shown, we require a lower standard of proof of substantial similarity. See Three Boys Music Corp. v. Bolton, 212 F.3d 477, 485 (9th Cir. 2000); Smith, 84 F.3d at 1218. For the purposes of summary judgment, Carey conceded that she had a high degree of access to One.[2] Swirsky's [*845] burden of proof of substantial similarity is thus commensurately lowered.
In determining whether two works are substantially similar, we employ a two-part analysis: an objective extrinsic test and a subjective intrinsic test. For the purposes of summary judgment, only the extrinsic test is important because the subjective question whether works are intrinsically similar must be left to the jury. See Rice I, 330 F.3d at 1174. If Swirsky cannot present evidence that would permit a trier of fact to find that he satisfied the extrinsic test, he necessarily loses on summary judgment because a "jury may not find substantial similarity without evidence on both the extrinsic and intrinsic tests." Rice I, 330 F.3d at 1174.
The extrinsic test considers whether two works share a similarity of ideas and expression as measured by external, objective criteria.[3]See Smith, 84 F.3d at 1218. The extrinsic test requires "analytical dissection of a work and expert testimony." Three Boys, 212 F.3d at 485. "Analytical dissection" requires breaking the works "down into their constituent elements, and comparing those elements for proof of copying as measured by 'substantial similarity.'" Rice v. Fox Broad. Co., 148 F. Supp. 2d 1029, 1051 (C.D. Cal. 2001), reversed on other grounds, 330 F.3d 1170 (9th Cir. 2003) (Rice II). Because the requirement is one of substantial similarity to protected elements of the copyrighted work, it is essential to distinguish between the protected and unprotected material in a plaintiff's work. See Rice I, 330 F.3d at 1174.
The expert testimony on which Swirsky relied was that of Dr. Robert Walser, chair of the Musicology Department at the University of California at Los Angeles. On the basis of his aural assessment[4]of One and Thank God, Dr. Walser opined that the two songs had substantially similar choruses.
Dr. Walser admitted that the lyrics and verse melodies of the two songs differed "clearly and significantly," but stated that the two songs' choruses shared a "basic shape and pitch emphasis" in their melodies, which were played over "highly similar basslines[5] and chord changes, at very nearly the same tempo and in the same generic style."[6] Dr. Walser also noted that it was a "suspicious coincidence" that the two songs' choruses were both sung in B-flat. Dr. Walser further testified that the choruses in both One and Thank God shared a similar structure in that measures five through seven of each chorus were "almost exactly" the same as the first three measures of each chorus.
[*846] Dr. Walser also noted a number of differences between the two songs' choruses. Dr. Walser found that the fourth measures of the choruses were "dramatically different" from each other and noted that while the "basic, emphasized pitches and rhythms" of the basslines were alike, the basslines to both choruses were "ornamented and played slightly differently from chorus to chorus." Dr. Walser also found that certain "text-setting choices"[7] created differences between the two songs' choruses. For example, he noted that in Thank God, Carey sings "D, scale degree three, for a full beat on the first beat of the first measure" while Xscape in One sings the same pitch "divided into two eight-note pulses." Dr. Walser ultimately concluded, however, that these differences were not enough to differentiate the songs because the overall emphasis on musical notes was the same, which "contributed to the impression of similarity one hears when comparing the two songs."
Dr. Walser transcribed his aural impressions into a series of visual "transcriptions." Dr. Walser created a transcription of each chorus' pitch sequence, melody,[8] and bassline. Dr. Walser labeled his transcription of the basslines a "reduction" because he transcribed only the "basic, emphasized pitches and rhythms." Dr. Walser thus did not include any bassline notes or pitches he found to be "ornamented" in his transcriptions.
The district court found this evidence insufficient to survive a motion for summary judgment for four reasons. First, the district court found that Dr. Walser's expert methodology was flawed. Second, the district court, using its own analysis, found that no triable issue was raised as to the substantial similarity of measures two, three, six, seven, and eight of the two choruses. Third, the district court held that measures one and five of One were scenes a faire,[9] and thus incapable of supporting a finding of infringement. Finally, the district court discounted any similarity between the two choruses based on key, harmony, tempo, or genre because it found no precedent for substantial similarity to be "founded solely on similarities in key, harmony, tempo or genre, either alone or in combination." We disagree with much of the district court's reasoning on all four points and conclude that Swirsky has satisfied the extrinsic test because he has provided "indicia of a sufficient disagreement concerning the substantial similarity of [the] two works."
A. Dr. Walser's Methodology
There is nothing inherently unsound about Dr. Walser's musicological methodology in this case. The district court is correct that Dr. Walser's methodology is "selective," inasmuch as it discounts notes that he characterizes as "ornamental." Dr. Walser, however, explained [*847] that the melody (pitch and rhythm) and bassline of a song cannot be divorced from the harmonic rhythm of a song. According to Dr. Walser, notes falling on the beat will be more prominent to the ear than notes falling off the beat. Thus, Dr. Walser opined that, even though measure three of both choruses were not identical in numerical pitch sequence or note selection, they both "emphasized the second scale degree, C, over an A in the bass, resolving to the third scale degree, D, over a D in the bass in the last half of the measure." Dr. Walser provided a comparable analysis for measures one, three, and eight.
Similarly, Dr. Walser explained that some artists will ornament their notes in ways that others do not. Dr. Walser testified at deposition that both Carey and Xscape ornament their notes with "melismas" and "appoggiaturas," both of which are technical terms for moving up to the next note and then back again. Dr. Walser testified that he did not notate these ornaments in his transcriptions, or take them into account in his opinion, because he "took that to be a matter of the singer customizing the song and regarded those notes as not structural; they are ornamental." As we said in Newton v. Diamond, 349 F.3d 591 (2003), we can "consider only [the defendant's] appropriation of the song's compositional elements and must remove from consideration all the elements unique to [Plaintiff's performance."Dr. Walser's methodology sought to remove notes he perceived as performance-related.
To a certain extent, Dr. Walser's methodology does concentrate on how the two choruses sound to his expert ears, which led the district court to conclude that his testimony related to intrinsic and not extrinsic similarity. We do not agree, however, that Dr. Walser's testimony was an intrinsic rather than extrinsic analysis. He was not testifying, as the intrinsic test would require, as to whether subjectively the "ordinary, reasonable person would find the total concept and feel of the [two choruses] to be substantially similar." Three Boys, 212 F.3d at 485. Instead, he was stating that, although the two choruses are not exactly identical on paper, when examined in the structural context of harmony, rhythm, and meter, they are remarkably similar. We, therefore, cannot accept the district court's conclusion that Dr. Walser did not "adequately explain, based on objective criteria, why [his] particular subset of notes is more important, or more appropriately analyzed, than the other notes present in the songs." The district court erred in completely discounting Dr. Walser's expert opinion.
B. The District Court's Measure-by-Measure Analysis
The district court also erred by basing its comparison of the two choruses almost entirely on a measure-by-measure comparison of melodic note sequences from the full transcriptions of the choruses.[10] Objective analysis of music under the extrinsic test cannot mean that a court may simply compare the numerical representations of pitch sequences and the visual representations of notes to determine that two choruses are not substantially similar, without regard to other elements [*848] of the compositions. Under that approach, expert testimony would not be required at all, for any person untrained in music could conclude that "2-2-2-2-2-2-1-2-1-3" did not match "2-2-4-3-2-3" or that a half-note is not identical to an eighth-note. Certainly, musicological experts can disagree as to whether an approach that highlights stressed notes, as Dr. Walser's does, is the most appropriate way to break down music for substantial-similarity comparison, but no approach can completely divorce pitch sequence and rhythm from harmonic chord progression, tempo, and key, and thereby support a conclusion that compositions are dissimilar as a matter of law. It is these elements that determine what notes and pitches are heard in a song and at what point in the song they are found. To pull these elements out of a song individually, without also looking at them in combination, is to perform an incomplete and distorted musicological analysis.[11]
Furthermore, to disregard chord progression, key, tempo, rhythm, and genre is to ignore the fact that a substantial similarity can be found in a combination of elements, even if those elements are individually unprotected. See Satava v. Lowry, 323 F.3d 805, 811 (9th Cir. 200). Thus, although chord progressions may not be individually protected, if in combination with rhythm and pitch sequence, they show the chorus of Thank God to be substantially similar to the chorus of One, infringement can be found. See Three Boys, 212 F.3d at 485; Satava, 323 F.3d at 811. [12]
We recognize the difficulties faced by the district court in this case. We have referred to "the turbid waters of the 'extrinsic test' for substantial similarity under the Copyright Act." Metcalf, 294 F.3d at 1071. The application of the extrinsic test, which assesses substantial similarity of ideas and expression, to musical compositions is a somewhat unnatural task, guided by relatively little precedent. Music is an art form that "produces sounds and expresses moods," Debra Presti Brent, The Successful Musical Copyright Infringement Suit: The Impossible Dream, 7 U. Miami Ent. & Sports. L. Rev. 229, 244 (1990), but it does not necessarily communicate separately identifiable ideas. The extrinsic test provides an awkward framework to apply to copyrighted works like music or art objects, which lack distinct elements of idea and expression. Nevertheless, the test is our law and we must apply it. The extrinsic test does serve the purpose of permitting summary judgment in clear cases of non-infringement, and it informs [*849] the fact-finder of some of the complexities of the medium in issue while guiding attention toward protected elements and away from unprotected elements of a composition.
In analyzing musical compositions under the extrinsic test, we have never announced a uniform set of factors to be used. We will not do so now. Music, like software programs and art objects, is not capable of ready classification into only five or six constituent elements; music is comprised of a large array of elements, some combination of which is protectable by copyright.[13] For example, in Three Boys we upheld a jury finding of substantial similarity based on the combination of five otherwise unprotectable elements: (1) the title hook phrase (including the lyric, rhythm, and pitch); (2) the shifted cadence; (3) the instrumental figures; (4) the verse/chorus relationship; and (5) the fade ending. Three Boys, 212 F.3d at 485. Other courts have taken account of additional components of musical compositions, including melody, harmony, rhythm, pitch, tempo, phrasing, structure, chord progressions, and lyrics. See Ellis v. Diffie, 177 F.3d 503, 506 (6th Cir. 1999) (noting that the district court had compared idea, phraseology, lyrics, rhythms, chord progressions, "melodic contours," structures, and melodies under "ordinary observer" test); Cottrill v. Spears, 2003 U.S. Dist. LEXIS 8823, 2003 WL 21223846, at *9 (E.D. Pa. May 22, 2003) (unpublished disposition) (comparing pitch, chord progression, meter, and lyrics under extrinsic test); Tisi v. Patrick, 97 F. Supp. 2d 539, 543 (S.D.N.Y. 2000) (analyzing structure, melody, harmony, and rhythm under "striking similarity" test); McKinley v. Raye, 1998 U.S. Dist. LEXIS 3019, 1998 WL 119540, at *5 (N.D. Tex. March 10, 1998) (mem.) (analyzing lyrics, melodies, and song structure); Damiano v. Sony Music Entm't, Inc., 975 F. Supp. 623, 631 (D. N.J. 1996) (analyzing instrumentation and melody under the extrinsic test); Sylvestre v. Oswald, 1993 U.S. Dist. LEXIS 7002, 1993 WL 179101, at *4 (S.D.N.Y. May 18, 1993) (analyzing melody and lyrics under "striking similarity" test); Intersong-USA v. CBS, Inc., 757 F. Supp. 274, 280 (S.D.N.Y. 1991) (analyzing chord progression, structure, pitch, and harmony under substantial similarity[14] test). In addition, commentators have opined that timbre, tone, spatial organization, consonance, dissonance, accents, note choice, combinations, interplay of instruments, basslines, and new technological sounds can all be elements of a musical composition. See Brent, supra, at 248-89; Stephanie J. Jones, Music Copyright in Theory and Practice: An Improved Approach for Determining Substantial Similarity, 31 Duq. L. Rev. 277, 294-95 (1993).
There is no one magical combination of these factors that will automatically substantiate a musical infringement suit; each allegation of infringement will be unique. So long as the plaintiff can demonstrate, through expert testimony that addresses some or all of these elements and supports its employment of them, that the similarity was "substantial" and to "protected elements" of the copyrighted work, the extrinsic test is satisfied. Swirsky has met that standard here.
C. Scenes a Faire Analysis
The district court erred in finding the first and fifth measures of One to [*850] be unprotectable by reason of the scenes a faire doctrine. Scenes a faire analysis requires the court to examine whether "motive"[15] similarities that plaintiffs attribute to copying could actually be explained by the common-place presence of the same or similar "motives" within the relevant field. See Smith, 84 F.3d at 1219. Under the scenes a faire doctrine, when certain commonplace expressions are indispensable and naturally associated with the treatment of a given idea, those expressions are treated like ideas and therefore not protected by copyright. See Rice I, 330 F.3d at 1175. The district court held that the first and fifth measures of One [**23] were not protected by copyright because Dr. Walser admitted in his deposition that the pitch sequence of the first measure of One's chorus was more similar to the pitch sequence in the first measure of the folk song "For He's a Jolly Good Fellow" ("Jolly Good") than to the pitch sequence in the first measure of Thank God's chorus.